About 567politics: AI-assisted news that examines every side

What 567politics is

567politics covers politics in the United States and around the world: elections, governments, policy fights, the people who hold office, and the money and data behind them. We use AI to research and write that coverage, and we hold it to one goal: help readers understand a political question from every serious side, using the best available evidence, without telling them which team to cheer for.

Much political coverage is written for one side's audience. We are trying to build something else: a publication that a progressive, a conservative, a libertarian, a socialist and someone who has never voted could all read and come away feeling that their own view was described fairly, having also learned the strongest case for the views they don't hold. Our slogan says it in three lines: Independent news. Every side examined. Sources and data linked.

Our commitments in brief

Neutral by design. The same rules apply to every party, ideology, country and public figure. We don't endorse candidates or tell anyone how to vote.
Every side's strongest case. We present each position the way its own supporters would, sourced to their own words. No straw men.
Primary sources and data first. Original documents, official statistics, court records, filings and representative polls, linked beside the claims they support.
Explanation over emotion. We lead with the question, the evidence and the tradeoffs, not with stories built to make you feel one way.
No partisan propaganda. No talking points repeated as fact, no loaded labels, no outrage headlines.
Global by default. Each country's politics is covered on its own terms, with domestic voices first.
Honest about limits. We make mistakes. We show our sources so you can check us, and we correct what we get wrong.

How we use AI

567politics is an AI-assisted publication. Our stories, politician profiles and issue guides are researched and drafted by AI working under the written editorial standards that govern everything we publish. We say so plainly because readers deserve to know how their news is made.

What the AI does

Reads the primary record: bills and laws, court filings, official statistics, election results, government reports, campaign-finance filings and the published statements of the people involved.
Pulls official data directly. For the United States that includes the Bureau of Labor Statistics, the Bureau of Economic Analysis, the Census Bureau, the Treasury, the FBI, the Bureau of Justice Statistics, Justice Department foreign-agent (FARA) filings and Federal Election Commission records. Each figure is cited with its observation period and the context needed to read it.
Searches deliberately for disagreement: separate searches for reporting, for arguments in support and for criticism and opposing arguments, so a story isn't built from one side's coverage.
Reads our earlier coverage, so new stories link to related reporting, profiles, races and explainers.
Drafts the story, its differing-perspectives section and its list of alternate sources, then checks the draft against our standards before it can be published.

Why AI

Used carefully, AI is good at the parts of political reporting that tend to get skipped on deadline: reading the whole bill instead of the press release about it, pulling the underlying dataset instead of quoting someone's chart, and reading the arguments of the side a writer disagrees with as closely as the side they agree with. It can apply the same checklist to every story, every party and every country, every time.

Why AI alone isn't enough

AI is not neutral by nature. Language models learn from human writing, which carries every bias people have, and they can state an error with confidence. So we don't rely on a model's instincts. We constrain it: written rules for evidence and fairness, required sections that make every side part of the story, automated checks a story must pass before it can be published, and links to sources so any reader can verify a claim. When the evidence for a claim is missing, the rule is to narrow the claim or hold the story, never to fill the gap with invented data, quotes, polls or posts.

Neutral by design

Neutrality is a design goal here, and we build it into how every piece is made. We try to be as neutral as possible and to respect the views of people on all sides.

One standard for everyone. The same evidence rules apply to every party, ideology, country and public figure. A claim from a politician readers might like gets checked as hard as one from a politician they might dislike.
No endorsements. We don't back parties or candidates, and we don't tell readers how to vote.
Facts labeled as facts. We keep verified facts, attributed claims, analysis, forecasts and opinion distinct, so you can tell which is which.
More than two sides. Few issues divide neatly between two parties. We include independents, minority viewpoints, the people most affected, and the disagreements inside each camp.
Respect for people who disagree. Political disagreements often come down to how people weigh values they share, such as safety, prosperity, fairness and freedom. We describe each view with the respect its holders would expect, and we don't speculate about anyone's motives.

Neutral does not mean pretending every claim is equally true. Giving arguments comparable attention doesn't require treating them as equally supported. When the evidence on a factual question is clear, we say so, by the same standard for every side. When a disagreement is about values, such as how to weigh liberty against security or growth against equality, we lay out the tradeoffs and leave the judgment to you.

No straw men, no propaganda

A straw man is a weak or distorted version of an argument, set up so it is easy to knock down. It is one of the most common tools in partisan media, and we work never to use it. Instead we aim to steelman: to present each position in its strongest form, in words its own supporters would recognize and accept as fair.

Each position is sourced to its own advocates, in their own words, not described secondhand by their opponents.
We explain each side's best evidence, the tradeoffs it accepts and the strongest objections to it.
Left-wing and right-wing arguments get comparable space, depth, prominence and scrutiny when both are relevant.
If a side has no verifiable argument or response on record, we say so rather than inventing one for it.
When an argument from any side contains a clear logical fallacy, we quote it, name the error and explain it. We don't use fallacy labels as insults, and a flawed argument doesn't by itself prove its conclusion wrong.

We also stay away from the building blocks of propaganda: a side's talking points repeated as settled fact, labels chosen to make one side look bad, numbers without denominators or time periods, quotes stripped of context, rumors presented as established, and headlines written to provoke outrage. Our publishing system rejects headlines and copy that lean on stock clickbait such as “shocking” or “you won't believe.”

Explanation, not heartstrings

Many political stories open with one sympathetic person, often a family, a worker or a patient, whose struggle stands in for an entire policy debate. Those stories can be moving, and the people in them are real. But one story can't tell you how common a problem is, what causes it, what a proposed policy would cost or who else it would affect. And every side of a debate can find a sympathetic face to put at the top of its version of events.

So we don't build stories around struggle narratives designed to pull at your heartstrings. We lead with the question and the evidence:

How many people are affected, out of how many, and over what period?
What does the best data show, and what are its limits?
What does each side say would happen under its approach, and what costs does it accept?
Who gains and who loses under each option?

Personal experiences still matter. When we include them, we present them as examples, not proof; we show how typical they are; and we include people affected on different sides of the issue. The aim is to understand each issue from all sides before deciding how to feel about it.

Every side, in every story

Two parts of every 567politics story exist to make sure no side is left out.

Differing perspectives

Every story carries a Differing perspectives section, and our publishing system won't publish a story without one. It states the disputed question in neutral terms, then sets out two to five substantive positions: who holds each one, its case as its own supporters would make it, and a link to their own words. Each position gets comparable length and identical presentation, and the positions appear in alphabetical order so that their order implies no ranking. Positions aren't graded inside the section; the story itself weighs the evidence, by the same standard for every side.

Alternate sources

Stories and politician profiles end with a reviewed list of alternate sources: independent reporting and primary evidence from other publishers, including ones that reach different conclusions than we do. Each link names its publisher, explains why it is relevant and carries the date it was reviewed. We look for at least two genuinely independent publishers when they exist, and copies of the same syndicated wire story don't count as independent confirmation.

Primary sources and data

We would rather show you the document than tell you what someone said about it.

Original records first. Laws and bills, court records, official datasets, government reports, election results, campaign filings and published research come before commentary about them.
Citations at the claim. Each factual claim or quotation links to its source right beside it, not in a list at the bottom.
Numbers with context. Figures come with their denominators, time periods and uncertainty, because a number without a comparison rarely means much.
Contested claims attributed. When something is disputed, we say who claims it and check it independently. Allegations stay separate from findings, and legal outcomes are stated accurately.
Campaign money, matched carefully. Politician profiles show Federal Election Commission data only for verified FEC candidate and committee IDs, never for a match guessed from a name. A company's employee PAC is reported as that, not as the company's own money.
Missing data said out loud. When data is unavailable, outdated or limited, we say so instead of guessing.

Polls, posts and markets

Claims about what most people think require credible representative polling. We cite the original poll with its field dates, population, sample size, question wording, sponsor and margin of error, and we distinguish adults, registered voters and likely voters. Social media posts can show how some people are reacting, and we quote them as that, but likes, replies and viral threads don't establish public opinion. Prediction market prices are shown market by market, never averaged, and they aren't polls or forecasts; our methodology explains how we collect them.

Global coverage

567politics covers elections and governments around the world, not only in the United States. Countries get an election guide, a watchlist of 20 issues that shape their politics, and profiles of their leading figures, and our country comparisons put population, income, purchasing power and economic data side by side.

We approach every country's politics from an outsider's perspective: independent of its partisan camps and curious about all of them. That describes our stance, not a claim that we have no point of view. When we describe how people in a country feel, we put domestic voices first and label foreign commentary, diaspora views, officials and organized advocacy separately. We don't assume where someone lives or what nationality they hold from their name, language or appearance. The same standards apply in every country.

What you'll find here

The newsroom: long-form reporting, 1,500 to 5,000 words per new story, each with differing perspectives and alternate sources.
Politicians: profiles with career timelines, sourced issue positions and, for U.S. federal candidates, campaign-finance data.
Issue guides: 20 issues for each country, each defined on a published scale.
Senate and governor races, the House map and President 2028: prediction market prices from each provider, shown separately with their rules.
World elections: country guides and international races.
Country compare: population, income, purchasing power and economic data for any two countries.
Political quiz: a private, statement-by-statement way to see where you stand. It is a tool for reflection, not a validated ideology test.
Data and tools: calculators for inflation, income percentiles and household finances.
Embeds: free maps, charts and calculators for other sites.

How issue positions are scored

Politician profiles place each person on defined scales for the issues in their country. A placement needs attributable evidence, such as a vote, a personal statement, an adopted platform or a government policy, recorded with its source, date and exact location. A party label alone doesn't count. When evidence shows only a range, we publish the range. When it doesn't establish a position, we say so instead of guessing a midpoint.

Checks before publication

Some of our standards are enforced by software, so a story that breaks them can't be published:

A Differing perspectives section with two to five positions is required, and a published story can't drop it.
New stories run 1,500 to 5,000 words, so complicated issues get room for every side.
Each new story links in context to at least four related 567politics pages, such as earlier reporting, profiles, races or explainers.
Clickbait phrases, placeholder text and all-caps headlines are rejected, and headlines stay under 110 characters.
Story text must read at or below a 12th-grade level on the Flesch-Kincaid scale, so analysis stays clear.

Other checks are part of the editorial standard every story is written to: claims and citations verified, competing views covered, polling kept separate from online reactions, loaded language removed and uncertainty disclosed. Drafts are also screened for formulaic AI phrasing, which is replaced with specific facts and direct sentences.

Our limits

No publication is free of bias, and we don't claim to be. Choosing which stories to cover is itself a judgment. AI can misread a document, miss context, or summarize one side more smoothly than another, and our standards guide the work rather than guarantee its accuracy. What we can promise is that we show our sources, apply the same rules to every side, and fix mistakes when we find them or you point them out.

Corrections and feedback

If you find an error, email founder@loungefrog.com with the page and, if you can, a source supporting the correction. We check it, fix it, and note material corrections in the story.

We especially want to hear when you think we have been unfair, to your side or to the other one. Tell us which argument we missed, misstated or presented in a weaker form than its supporters would. That feedback makes the coverage better.

Who runs 567politics

567politics is published by the team behind Lounge Frog, founded by Patrick Karsh, alongside the sister sites Points Frog, Cruises Frog and Seats Frog. One account works across all of them; our privacy policy explains what it holds. 567politics is independent: it isn't run by, or affiliated with, any political party, campaign, government or advocacy group.

You are welcome to quote and cite our reporting with a link to the page. Search engines and AI assistants may read the site, and llms.txt gives them a map of it.

Common questions

Is 567politics written by AI?

Yes. AI researches and drafts 567politics stories, profiles and issue guides under written editorial standards. Stories must pass automated checks before they can be published, and every claim links to its source so readers can check it.

Is 567politics left-leaning or right-leaning?

Neither, by design. 567politics applies the same evidence rules to every party and ideology, presents each side's strongest case in terms its own supporters would recognize, and endorses no party or candidate. No publication is perfectly free of bias, so readers who see an imbalance are asked to point it out.

Does 567politics only cover the United States?

No. 567politics covers elections, governments and policy debates around the world, with country election guides, 20-issue watchlists for each country, politician profiles and country-to-country data comparisons.

Where does 567politics get its information?

From primary sources first: laws, court records, official statistics, government reports, election results, campaign-finance filings, representative polls and the published words of the people involved. Each claim links to its source next to the text.

Why doesn't 567politics lead stories with personal struggle stories?

One person's story can't show how common a problem is, what causes it or what a policy would change, and every side of a debate can find a sympathetic example. 567politics leads with the question, the data and each side's arguments, and presents personal experiences as examples rather than proof.

How do I get 567politics by email?

Sign in with a free 567politics account from the daily email sign-up, on the About page or at the foot of any other page, and the email starts. It lists the day's new stories as headlines, comes once a day at most and never on quiet days, and every email has a one-click unsubscribe.

How do I report an error or an unfair story?

Email founder@loungefrog.com with the page and, if you can, a source supporting the correction. Material corrections are noted in the story.

Last updated November 2, 2026.